Set up

Import Data

Tables by Severity

Demographic

The table below provides demographic information based on severity of injury

# level naming for categorical variables
df_demo$gender <- factor(df_demo$gender,
                   levels = c(1,2,3),
                   labels = c("Male", "Female", "Nonbinary"))

df_demo$work_current <- factor(df_demo$work_current,
                   levels = c(1,0),
                   labels = c("Yes", "No"))

df_demo$severity <- factor(df_demo$severity,
                   levels = c(2,3),
                   labels = c("Moderate", "Severe"))

df_demo$mech_injury <- factor(df_demo$mech_injury,
                   levels = c(1,2,3,4,5),
                   labels = c("Fall", "MVC", "Sports", "Violence", "Pedestrian struck"))

df_demo$income <- factor(df_demo$income,
                   levels = c(1,2,3),
                   labels = c("<52K", "52K-156K", ">156K"))

df_demo$marital_status <- factor(df_demo$marital_status,
                                 levels = c(1, 2, 3, 4),
                                 labels = c("Single", "Married", "Divorced", "Widowed"))
Characteristic N Moderate, N = 191 Severe, N = 281 p-value2
Age (years) 47 51 (14) 43 (15) 0.062
Time since TBI (years) 47 7 (5) 10 (8) 0.17
Gender 47

0.17
    Male
7 (37%) 17 (61%)
    Female
11 (58%) 11 (39%)
    Nonbinary
1 (5.3%) 0 (0%)
Education (years) 47 15.47 (2.04) 15.00 (2.62) 0.49
Race/Ethnicity 47

0.25
    Asian
0 (0%) 2 (7.1%)
    Biracial
2 (11%) 0 (0%)
    Black
0 (0%) 1 (3.6%)
    Hispanic
1 (5.3%) 3 (11%)
    White
16 (84%) 22 (79%)
Employment status 47

0.62
    Yes
9 (47%) 10 (36%)
    No
10 (53%) 18 (64%)
Annual household income 47

0.83
    <52K
6 (32%) 10 (36%)
    52K-156K
9 (47%) 14 (50%)
    >156K
4 (21%) 4 (14%)
Size household 47 2.00 (1.05) 2.25 (1.38) 0.49
Marital status 47

0.089
    Single
5 (26%) 16 (57%)
    Married
11 (58%) 8 (29%)
    Divorced
3 (16%) 4 (14%)
    Widowed
0 (0%) 0 (0%)
Substance use score 47 4.16 (3.62) 1.75 (1.94) 0.014
Cause of injury 47

0.10
    Fall
10 (53%) 5 (18%)
    MVC
4 (21%) 12 (43%)
    Sports
1 (5.3%) 4 (14%)
    Violence
1 (5.3%) 4 (14%)
    Pedestrian struck
3 (16%) 3 (11%)
1 Mean (SD); n (%)
2 Welch Two Sample t-test; Pearson’s Chi-squared test

ACS

ACS3 (activity re-engagement scores - outcome measure) by severity of injury

Characteristic N Moderate, N = 191 Severe, N = 281 p-value2
ACS Global Before 47 72 (11) 68 (10) 0.23
ACS Global Current 47 54 (17) 51 (13) 0.53
Global Retained (%) 47 75 (19) 75 (16) 0.92
ACS IADL Before 47 22.16 (2.27) 21.04 (3.25) 0.17
ACS IADL Current 47 17.9 (4.9) 16.7 (4.4) 0.37
IADL Retained (%) 47 81 (19) 80 (18) 0.87
ACS Leisure Before 47 22.5 (5.8) 21.0 (4.5) 0.36
ACS Leisure Current 47 18.0 (6.0) 16.3 (5.1) 0.33
Leisure Retained (%) 47 82 (22) 78 (18) 0.58
ACS Fitness Before 47 13.2 (4.4) 13.0 (4.5) 0.89
ACS Fitness Current 47 8.3 (4.7) 8.3 (3.3)
0.99
Fitness Retained (%) 47 64 (32) 68 (35) 0.65
ACS Social Before 47 13.95 (1.22) 12.93 (1.54) 0.015
ACS Social Current 47 9.68 (3.08) 9.71 (2.64) 0.97
Social Retained (%) 47 69 (19) 75 (20) 0.27
1 Mean (SD)
2 Welch Two Sample t-test

Below is the ttest for the specific t and p value for the difference between ACS3 previous social score, which was significantly different.

## 
##  Welch Two Sample t-test
## 
## data:  df_mod$acss_prev and df_severe$acss_prev
## t = 2.5214, df = 43.755, p-value = 0.01541
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  0.2043487 1.8332453
## sample estimates:
## mean of x mean of y 
##  13.94737  12.92857

FrSBe

Comparison of self-regulation scores by severity

Characteristic N Moderate, N = 191 Severe, N = 281 p-value2
Executive function 47 41 (10) 43 (11) 0.67
Disinhibition 47 32.3 (6.1) 33.6 (6.5) 0.48
Apathy 47 33 (8) 34 (9) 0.57
Total FrSBe Score 47 106 (19) 110 (22) 0.50
1 Mean (SD)
2 Welch Two Sample t-test

TBI QOL

Comparison of subscales of TBI QOL measure by severity of injury

Characteristic N Moderate, N = 191 Severe, N = 281 p-value2
Participation SRA 47 47 (7) 46 (6) 0.44
Anger 47 53 (9) 50 (10) 0.34
Anxiety 47 58 (7) 54 (10) 0.14
Communication 47 46 (9) 46 (10) 0.85
Depression 47 55 (8) 53 (11) 0.60
Dyscontrol 47 52 (8) 50 (9) 0.53
EF 47 34.9 (4.8) 35.8 (6.8) 0.59
Fatigue 47 57 (8) 54 (9) 0.32
Gen Cognition 47 36 (8) 36 (9) 0.77
Headache 47 51 (9) 49 (9) 0.48
Mobility 47 47 (10) 44 (7) 0.25
Pain 47 58 (11) 54 (10) 0.24
Positive Effect 47 50 (6) 50 (8) 0.87
Resilience 47 49 (6) 48 (9) 0.54
Satisfaction SRA 47 46 (6) 45 (6) 0.50
Self esteem 47 47 (11) 48 (11) 0.74
Stigma 46 50 (8) 52 (7) 0.59
Upper Extremity 47 47 (9) 42 (8) 0.063
1 Mean (SD)
2 Welch Two Sample t-test

TBI Composite

Comparison of composite scores for TBI QOL by severity of injury. Composite scores were calculated using:

Tyner, C. E., Boulton, A. J., Sherer, M., Kisala, P. A., Glutting, J. J., & Tulsky, D. S. (2020). Development of Composite Scores for the TBI-QOL. Arch Phys Med Rehabil, 101(1), 43-53. https://doi.org/10.1016/j.apmr.2018.05.036

Characteristic N Moderate, N = 191 Severe, N = 281 p-value2
Physical Health Index 47 91 (14) 96 (14) 0.22
Emotional Health Index 47 97 (12) 101 (15) 0.29
Cognitive Health Index 47 93 (13) 95 (16) 0.77
Social Health Index 47 94 (12) 91 (12) 0.41
Global Health Index 47 93 (13) 95 (14) 0.55
1 Mean (SD)
2 Welch Two Sample t-test

Predictive variables

Table 2 in dissertation

This table compares only the Personal and Environmental Protective factors and self-regulation outlined in the dissertation. Note that the Cognitive Health Composite score was not used as it includes executive functioning, which in this paper is considered a self-regulatory process. Therefore, general cognitive functioning was used which assesses memory and concentration.

Characteristic N Moderate, N = 191 Severe, N = 281 p-value2
Physical Health Index 47 91 (14) 96 (14) 0.22
Emotional Health Index 47 97 (12) 101 (15) 0.29
General Cognition 47 36 (8) 36 (9) 0.77
Extraversion 47 7.16 (2.50) 6.89 (2.33) 0.72
Agreeable 47 7.11 (1.94) 7.11 (2.08)
0.99
Consciousness 47 8.16 (1.54) 7.75 (1.94) 0.43
Neuroticism 47 6.47 (2.20) 6.21 (2.63) 0.72
Openness 47 8.53 (2.09) 7.25 (1.94) 0.041
Annual household income 47

0.83
    <52K
6 (32%) 10 (36%)
    52K-156K
9 (47%) 14 (50%)
    >156K
4 (21%) 4 (14%)
Marital status 47

0.089
    Single
5 (26%) 16 (57%)
    Married
11 (58%) 8 (29%)
    Divorced
3 (16%) 4 (14%)
    Widowed
0 (0%) 0 (0%)
Social Support 47 84 (10) 76 (11) 0.014
Executive function 47 41 (10) 43 (11) 0.67
Disinhibition 47 32.3 (6.1) 33.6 (6.5) 0.48
Apathy 47 33 (8) 34 (9) 0.57
Total score 47 106 (19) 110 (22) 0.50
1 Mean (SD); n (%)
2 Welch Two Sample t-test; Pearson’s Chi-squared test
## 
##  Welch Two Sample t-test
## 
## data:  df_mod$bfi_openness and df_severe$bfi_openness
## t = 2.115, df = 36.744, p-value = 0.04127
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  0.05332003 2.49931155
## sample estimates:
## mean of x mean of y 
##  8.526316  7.250000

Below is the t-test for SPS total, which was significantly different between severity of injury

## 
##  Welch Two Sample t-test
## 
## data:  df_mod$spstotal and df_severe$spstotal
## t = 2.5765, df = 41.069, p-value = 0.01367
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##   1.737685 14.337503
## sample estimates:
## mean of x mean of y 
##  83.89474  75.85714

Descriptive Statistics

Descriptive statistics for each variable of interest for the data set including mean, median, SD, and IQR, kurtosis and se

vars n mean sd median trimmed mad min max range skew kurtosis se
record_id* 1 47 24.00 13.71 24.0 24.00 17.79 1.0 47.0 46.0 0.00 -1.28 2.00
age_current 2 47 46.51 14.77 45.0 46.44 17.79 21.0 72.0 51.0 0.15 -1.16 2.16
age_injury 3 47 37.62 15.07 34.0 36.87 16.31 18.0 66.0 48.0 0.38 -1.25 2.20
time_injury 4 47 8.96 7.08 7.0 8.06 5.93 1.0 30.0 29.0 1.16 0.79 1.03
gender 5 47 1.51 0.55 1.0 1.49 0.00 1.0 3.0 2.0 0.35 -1.14 0.08
race* 6 47 4.57 1.04 5.0 4.85 0.00 1.0 5.0 4.0 -2.43 4.76 0.15
edu 7 47 15.19 2.39 16.0 15.21 2.97 10.0 20.0 10.0 -0.15 -0.76 0.35
work_current 8 47 0.40 0.50 0.0 0.38 0.00 0.0 1.0 1.0 0.38 -1.90 0.07
hours_work 9 19 30.50 13.04 40.0 31.50 0.00 4.0 40.0 36.0 -0.74 -1.19 2.99
occ_years_pos 10 18 7.43 9.21 3.5 6.47 5.04 0.1 30.0 29.9 1.47 1.14 2.17
diff_occ 11 19 0.53 0.51 1.0 0.53 0.00 0.0 1.0 1.0 -0.10 -2.09 0.12
no_occ_stat 12 28 3.18 0.72 3.0 3.12 0.00 2.0 5.0 3.0 1.45 1.89 0.14
income 13 47 1.83 0.70 2.0 1.79 1.48 1.0 3.0 2.0 0.23 -1.01 0.10
house_size 14 47 2.15 1.25 2.0 1.97 1.48 1.0 7.0 6.0 1.55 3.07 0.18
children 15 47 0.43 0.50 0.0 0.41 0.00 0.0 1.0 1.0 0.29 -1.96 0.07
num_child 16 20 2.00 1.26 2.0 1.75 1.48 1.0 5.0 4.0 1.21 0.49 0.28
severity 17 47 2.60 0.50 3.0 2.62 0.00 2.0 3.0 1.0 -0.38 -1.90 0.07
mech_injury 18 47 2.38 1.38 2.0 2.26 1.48 1.0 5.0 4.0 0.73 -0.80 0.20
mech_injury_other* 19 6 1.17 0.41 1.0 1.17 0.00 1.0 2.0 1.0 1.36 -0.08 0.17
substance 20 47 2.72 2.96 2.0 2.28 2.97 0.0 13.0 13.0 1.31 1.77 0.43
acsg_prev 21 47 69.55 10.21 68.0 69.38 10.38 48.0 93.0 45.0 0.23 -0.67 1.49
acsg_curr 22 47 52.16 14.35 49.0 51.28 14.53 29.5 85.5 56.0 0.53 -0.44 2.09
acsg_retain 23 47 75.21 17.11 74.0 75.46 14.83 40.0 107.0 67.0 -0.14 -0.48 2.50
acsi_prev 24 47 21.49 2.92 21.0 21.59 2.97 12.0 26.0 14.0 -0.53 0.62 0.43
acsi_curr 25 47 17.17 4.60 17.0 16.87 5.93 10.5 26.0 15.5 0.45 -1.05 0.67
acsi_retain 26 47 80.17 18.37 82.0 80.64 22.24 42.0 110.0 68.0 -0.16 -1.13 2.68
acsl_prev 27 47 21.60 5.05 21.0 21.64 5.93 11.0 32.0 21.0 -0.04 -0.82 0.74
acsl_curr 28 47 17.01 5.46 16.0 16.76 5.93 8.5 28.0 19.5 0.42 -0.99 0.80
acsl_retain 29 47 79.57 19.28 78.0 79.90 17.79 38.0 122.0 84.0 -0.11 -0.49 2.81
acsf_prev 30 47 13.11 4.38 13.0 13.23 4.45 4.0 20.0 16.0 -0.24 -0.96 0.64
acsf_curr 31 47 8.27 3.91 8.0 8.08 3.71 1.0 18.0 17.0 0.35 -0.09 0.57
acsf_retain 32 47 66.19 33.29 63.0 62.90 23.72 9.0 200.0 191.0 1.53 3.93 4.86
acss_prev 33 47 13.34 1.49 14.0 13.44 1.48 9.0 17.0 8.0 -0.46 0.51 0.22
acss_curr 34 47 9.70 2.79 10.0 9.77 2.97 4.5 15.0 10.5 -0.26 -0.70 0.41
acss_retain 35 47 72.79 19.57 73.0 73.51 19.27 32.0 109.0 77.0 -0.24 -0.82 2.85
activity_card_sort_complete 36 47 2.00 0.00 2.0 2.00 0.00 2.0 2.0 0.0 NaN NaN 0.00
spstotal 37 47 79.11 11.29 82.0 79.56 14.83 55.0 96.0 41.0 -0.28 -1.23 1.65
bfi_extraversion 38 47 7.00 2.38 7.0 7.13 2.97 2.0 10.0 8.0 -0.26 -1.03 0.35
bfi_agreeable 39 47 7.11 2.00 7.0 7.18 2.97 3.0 10.0 7.0 -0.37 -0.92 0.29
bfi_consciousness 40 47 7.91 1.78 8.0 8.08 1.48 3.0 10.0 7.0 -0.71 -0.32 0.26
bfi_neuroticism 41 47 6.32 2.44 6.0 6.36 2.97 2.0 10.0 8.0 -0.06 -1.14 0.36
bfi_openness 42 47 7.77 2.08 8.0 7.95 2.97 2.0 10.0 8.0 -0.51 -0.56 0.30
frsbe_exec 43 47 42.19 10.29 43.0 41.97 10.38 24.0 63.0 39.0 0.10 -0.89 1.50
frsbe_apathy 44 47 33.36 8.53 32.0 32.92 10.38 18.0 53.0 35.0 0.41 -0.70 1.24
frsbe_disinhib 45 47 33.06 6.35 32.0 32.82 7.41 21.0 49.0 28.0 0.34 -0.48 0.93
frsbe_total 46 47 108.62 20.68 109.0 108.18 23.72 72.0 150.0 78.0 0.20 -0.95 3.02
frsbe_complete 47 47 2.00 0.00 2.0 2.00 0.00 2.0 2.0 0.0 NaN NaN 0.00
tbiqol_part_sra_tscore 48 47 46.25 6.66 45.9 45.72 5.78 32.1 64.1 32.0 0.87 1.17 0.97
tbiqol_anger_tscore 49 47 51.32 9.92 51.7 51.27 11.56 33.1 69.9 36.8 0.03 -1.05 1.45
tbiqol_anxiety_tscore 50 47 55.69 9.14 56.8 55.95 10.08 36.1 73.0 36.9 -0.25 -0.68 1.33
tbiqol_comm_tscore 51 47 46.04 9.56 44.9 45.89 9.49 29.2 65.5 36.3 0.19 -0.88 1.39
tbiqol_depression_tscore 52 47 54.03 9.69 53.9 54.19 10.08 33.6 74.0 40.4 -0.11 -0.68 1.41
tbiqol_dyscontrol_tscore 53 47 50.99 8.21 52.4 51.42 7.86 33.2 66.8 33.6 -0.45 -0.45 1.20
tbiqol_execfunc_tscore 54 47 35.44 6.00 34.3 35.17 5.19 24.3 50.8 26.5 0.41 -0.46 0.88
tbiqol_fatigue_tscore 55 47 55.09 8.52 54.7 55.18 8.15 37.9 72.5 34.6 -0.02 -0.67 1.24
tbiqol_genconcern_tscore 56 47 36.09 8.72 35.7 35.91 8.60 19.7 53.8 34.1 0.21 -0.67 1.27
tbiqol_grief_tscore 57 47 52.78 9.45 53.9 53.35 8.01 30.7 70.3 39.6 -0.62 -0.13 1.38
tbiqol_headache_tscore 58 47 49.63 9.22 49.5 49.23 13.64 38.5 67.1 28.6 0.10 -1.36 1.34
tbiqol_mobility_tscore 59 47 45.56 8.70 44.1 45.12 8.01 31.5 63.6 32.1 0.49 -0.68 1.27
tbiqol_pain_tscore 60 47 55.41 10.67 57.4 55.42 11.12 38.4 74.8 36.4 -0.28 -1.06 1.56
tbiqol_posaffect_tscore 61 47 49.87 7.51 49.3 49.72 7.86 35.4 68.9 33.5 0.22 -0.56 1.10
tbiqol_resilience_tscore 62 47 48.41 8.01 49.1 48.09 7.41 33.4 73.6 40.2 0.46 0.51 1.17
tbiqol_selfesteem_tscore 63 47 47.93 10.65 48.0 47.94 10.82 28.4 66.0 37.6 0.02 -0.95 1.55
tbiqol_satissra_tscore 64 47 45.25 6.19 45.1 44.82 4.60 34.7 63.2 28.5 0.88 1.27 0.90
tbiqol_stigma_tscore 65 46 51.04 7.32 52.0 51.67 5.63 33.5 62.3 28.8 -0.76 -0.05 1.08
tbiqol_ue_tscore 66 47 43.97 8.55 42.5 43.83 8.30 27.9 58.1 30.2 0.43 -0.90 1.25
marital_status 67 47 1.70 0.72 2.0 1.64 1.48 1.0 3.0 2.0 0.49 -1.01 0.10
phys_health 68 47 110.50 17.08 110.9 110.54 19.87 82.6 143.3 60.7 -0.10 -1.10 2.49
phys_health_index 69 47 94.36 13.77 95.0 94.49 16.31 64.0 117.0 53.0 -0.08 -0.97 2.01
emo_health 70 47 161.04 25.49 158.9 161.67 31.73 115.1 201.3 86.2 -0.13 -1.25 3.72
emo_health_index 71 47 99.15 14.00 101.0 98.90 17.79 77.0 123.0 46.0 0.08 -1.33 2.04
cog_health 72 47 71.53 14.22 70.0 71.19 13.94 44.0 104.6 60.6 0.30 -0.56 2.07
cog_health_index 73 47 94.00 14.59 92.0 94.00 14.83 62.0 123.0 61.0 0.04 -0.71 2.13
soc_health 74 47 91.50 11.93 90.0 90.83 10.38 69.6 127.3 57.7 0.69 0.67 1.74
soc_health_index 75 47 92.64 12.10 92.0 92.87 10.38 64.0 122.0 58.0 -0.16 0.34 1.76
glob_health 76 47 380.15 45.21 380.0 379.26 54.86 303.0 468.0 165.0 0.11 -1.11 6.59
glob_health_index 77 47 94.23 13.32 95.0 93.95 14.83 71.0 120.0 49.0 0.06 -1.03 1.94

ACS3

Outcome variable: ACS3 for all scores mean(sd)

Characteristic N = 471
ACS Global Before 70 (10)
ACS Global Current 52 (14)
Global Retained (%) 75 (17)
ACS IADL Before 21.49 (2.92)
ACS IADL Current 17.2 (4.6)
IADL Retained (%) 80 (18)
ACS Leisure Before 21.6 (5.1)
ACS Leisure Current 17.0 (5.5)
Leisure Retained (%) 80 (19)
ACS Fitness Before 13.1 (4.4)
ACS Fitness Current 8.3 (3.9)
Fitness Retained (%) 66 (33)
ACS Social Before 13.34 (1.49)
ACS Social Current 9.70 (2.79)
Social Retained (%) 73 (20)
1 Mean (SD)

Correlations

All variables

Correlation of all variables of interest with TBI QOL subscores. While too small to read in HTML print out, nice reference during analysis

This is pretty hard to read, so the following matrices break it down into smaller parts

Correlation matrix of PPF only (using composite TBIQOL Scores)

Included Variables

Matrix with heat map for all included variables in dissertation. Figure 4 in dissertation

Below is the breakdown of all TBIQOL sub scores with the ACS3. While not included in this study, helpful for discussion and future publications.

FrSBe and TBI-QOL

TBI QOL subscales with the FrSBe. I think this could be really interested as a future paper given how some subscales of the TBI QOL overlap with FrSBe (Exec functioning, anger, dyscontrol)

RQ1: Regression Analysis

Research Question 1 1. What is the relationship between protective factors and self-regulation with resiliency-related outcomes such as re-engagement in meaningful activities? a. To what extent do protective factors and self-regulation predict resiliency-related outcomes in the TBI population? Hypothesis: Higher self-regulation will be associated with better resiliency-related outcomes b. To what extent does self-regulation mediate or moderate the influence of protective factors on resiliency-related outcomes after TBI? Hypothesis: Self-regulation will impact the relationship between protective factors and resiliency-related outcomes

RQ1a

First, we’ll look at the hierarchical linear model as outlined in Chapter 3. Then, to dive deeper, a “post hoc” analysis of each subscale of the ACS and use AIC to determine model of best fit.

ACS Global

In this section, we’ll do the original hierarchical model with protective and environmental protective factors in the first step and then total self-regulation score added for the second. *note that the cognitive composite score is not included as it includes exec functioning, which in this paper is seen as a self-regulatory process. therefore, gen concerns (memory and concentration) is used as cognitive protective factor

step1 <- lm(acsg_retain~age_current, data=df)
summary(step1)
## 
## Call:
## lm(formula = acsg_retain ~ age_current, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -41.363 -10.024  -0.443   9.097  35.486 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  86.4267     8.2323  10.498 1.11e-13 ***
## age_current  -0.2411     0.1689  -1.428     0.16    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 16.92 on 45 degrees of freedom
## Multiple R-squared:  0.04334,    Adjusted R-squared:  0.02208 
## F-statistic: 2.039 on 1 and 45 DF,  p-value: 0.1602
step2<- lm(acsg_retain~age_current+phys_health_index+tbiqol_genconcern_tscore+emo_health_index+ spstotal, data=df)
summary(step2)
## 
## Call:
## lm(formula = acsg_retain ~ age_current + phys_health_index + 
##     tbiqol_genconcern_tscore + emo_health_index + spstotal, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -29.849  -9.614  -1.017  10.909  33.458 
## 
## Coefficients:
##                          Estimate Std. Error t value Pr(>|t|)  
## (Intercept)              24.58124   25.56886   0.961   0.3420  
## age_current              -0.01410    0.16700  -0.084   0.9331  
## phys_health_index         0.14439    0.23883   0.605   0.5488  
## tbiqol_genconcern_tscore  0.79509    0.38979   2.040   0.0478 *
## emo_health_index         -0.09938    0.23424  -0.424   0.6736  
## spstotal                  0.23788    0.25473   0.934   0.3558  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 15.38 on 41 degrees of freedom
## Multiple R-squared:  0.2803, Adjusted R-squared:  0.1926 
## F-statistic: 3.194 on 5 and 41 DF,  p-value: 0.01586
#Nested Model Comparison
anova(step1, step2)
## Analysis of Variance Table
## 
## Model 1: acsg_retain ~ age_current
## Model 2: acsg_retain ~ age_current + phys_health_index + tbiqol_genconcern_tscore + 
##     emo_health_index + spstotal
##   Res.Df     RSS Df Sum of Sq      F  Pr(>F)  
## 1     45 12884.1                              
## 2     41  9692.2  4    3191.9 3.3756 0.01776 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#change in R-squared
summary(step2)$r.squared - summary(step1)$r.squared
## [1] 0.2370005
step3<- lm(acsg_retain~age_current+phys_health_index+tbiqol_genconcern_tscore+emo_health_index+ spstotal+frsbe_total, data=df)
summary(step3)
## 
## Call:
## lm(formula = acsg_retain ~ age_current + phys_health_index + 
##     tbiqol_genconcern_tscore + emo_health_index + spstotal + 
##     frsbe_total, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -27.554  -9.303   0.821   9.139  33.358 
## 
## Coefficients:
##                          Estimate Std. Error t value Pr(>|t|)
## (Intercept)              51.19175   40.65439   1.259    0.215
## age_current              -0.03531    0.16946  -0.208    0.836
## phys_health_index         0.14505    0.23967   0.605    0.548
## tbiqol_genconcern_tscore  0.67895    0.41467   1.637    0.109
## emo_health_index         -0.13345    0.23851  -0.560    0.579
## spstotal                  0.17925    0.26491   0.677    0.503
## frsbe_total              -0.12408    0.14704  -0.844    0.404
## 
## Residual standard error: 15.43 on 40 degrees of freedom
## Multiple R-squared:  0.2929, Adjusted R-squared:  0.1869 
## F-statistic: 2.762 on 6 and 40 DF,  p-value: 0.02427
#Nested Model Comparison
anova(step2, step3)
## Analysis of Variance Table
## 
## Model 1: acsg_retain ~ age_current + phys_health_index + tbiqol_genconcern_tscore + 
##     emo_health_index + spstotal
## Model 2: acsg_retain ~ age_current + phys_health_index + tbiqol_genconcern_tscore + 
##     emo_health_index + spstotal + frsbe_total
##   Res.Df    RSS Df Sum of Sq      F Pr(>F)
## 1     41 9692.2                           
## 2     40 9522.7  1    169.53 0.7121 0.4038
#change in R-squared
summary(step3)$r.squared - summary(step2)$r.squared
## [1] 0.01258785

Assumptions tested

Test for multicollinearity

##              age_current        phys_health_index tbiqol_genconcern_tscore 
##                 1.211297                 2.104684                 2.526230 
##         emo_health_index                 spstotal              frsbe_total 
##                 2.153464                 1.729262                 1.786663

Post Hoc AIC Models

As we have a smaller n and need to be parsimonious with the variables we use in the regression model, we’ll look at several models based on correlations (higher correlations added to the models) and then calculate the AIC. The lower AIC, the better the fit and that model will be used

Note that all assumptions were tested for each model and were met

ACS3 Social

## 
## Model selection based on AICc:
## 
##         K   AICc Delta_AICc AICcWt Cum.Wt      LL
## model7  7 403.69       0.00   0.24   0.24 -193.41
## model11 8 404.09       0.40   0.19   0.43 -192.15
## model9  7 404.37       0.69   0.17   0.60 -193.75
## model10 7 404.37       0.69   0.17   0.77 -193.75
## model6  8 406.45       2.77   0.06   0.83 -193.33
## model8  7 406.86       3.18   0.05   0.88 -195.00
## model1  6 407.72       4.04   0.03   0.91 -196.81
## model3  7 407.79       4.10   0.03   0.94 -195.46
## model4  7 407.82       4.14   0.03   0.97 -195.48
## model2  7 409.32       5.63   0.01   0.99 -196.22
## model5  8 409.42       5.73   0.01   1.00 -194.81
## 
## Call:
## lm(formula = acss_retain ~ tbiqol_fatigue_tscore + tbiqol_genconcern_tscore + 
##     tbiqol_depression_tscore + tbiqol_anxiety_tscore + frsbe_apathy, 
##     data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -35.235  -9.252   0.203  13.041  24.527 
## 
## Coefficients:
##                           Estimate Std. Error t value Pr(>|t|)   
## (Intercept)              102.31060   32.70059   3.129  0.00323 **
## tbiqol_fatigue_tscore      0.01614    0.39028   0.041  0.96722   
## tbiqol_genconcern_tscore   0.63535    0.36082   1.761  0.08572 . 
## tbiqol_depression_tscore   0.36486    0.37366   0.976  0.33457   
## tbiqol_anxiety_tscore     -0.86993    0.36947  -2.355  0.02342 * 
## frsbe_apathy              -0.73755    0.34611  -2.131  0.03913 * 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 15.87 on 41 degrees of freedom
## Multiple R-squared:  0.414,  Adjusted R-squared:  0.3425 
## F-statistic: 5.792 on 5 and 41 DF,  p-value: 0.0003912

ACS3 IADL

## 
## Model selection based on AICc:
## 
##        K   AICc Delta_AICc AICcWt Cum.Wt      LL
## model6 7 405.58       0.00   0.33   0.33 -194.35
## model4 7 405.83       0.25   0.29   0.61 -194.48
## model3 8 407.08       1.50   0.15   0.77 -193.64
## model5 8 408.23       2.65   0.09   0.85 -194.22
## model2 8 408.54       2.96   0.07   0.93 -194.37
## model1 8 408.60       3.02   0.07   1.00 -194.41
## 
## Call:
## lm(formula = acsi_retain ~ tbiqol_mobility_tscore + tbiqol_genconcern_tscore + 
##     tbiqol_comm_tscore + tbiqol_ue_tscore + frsbe_apathy, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -34.856 -10.388  -0.861  12.723  29.554 
## 
## Coefficients:
##                          Estimate Std. Error t value Pr(>|t|)  
## (Intercept)               35.5842    22.9859   1.548   0.1293  
## tbiqol_mobility_tscore     0.5331     0.4564   1.168   0.2495  
## tbiqol_genconcern_tscore   0.6523     0.3581   1.822   0.0758 .
## tbiqol_comm_tscore        -0.2707     0.3880  -0.698   0.4893  
## tbiqol_ue_tscore           0.3729     0.4641   0.803   0.4263  
## frsbe_apathy              -0.2153     0.3294  -0.654   0.5170  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 16.19 on 41 degrees of freedom
## Multiple R-squared:  0.3072, Adjusted R-squared:  0.2228 
## F-statistic: 3.637 on 5 and 41 DF,  p-value: 0.008165

ACS3 Leisure

## 
## Model selection based on AICc:
## 
##        K   AICc Delta_AICc AICcWt Cum.Wt      LL
## model7 6 409.52       0.00   0.34   0.34 -197.71
## model3 6 410.03       0.51   0.26   0.60 -197.97
## model6 6 410.45       0.93   0.21   0.81 -198.17
## model1 7 412.27       2.75   0.09   0.89 -197.70
## model4 7 412.64       3.12   0.07   0.97 -197.88
## model2 8 415.14       5.61   0.02   0.99 -197.67
## model5 7 415.84       6.32   0.01   1.00 -199.48
## 
## Call:
## lm(formula = acsl_retain ~ tbiqol_mobility_tscore + tbiqol_genconcern_tscore + 
##     tbiqol_comm_tscore + frsbe_apathy, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -36.799 -10.757  -0.328  10.936  36.487 
## 
## Coefficients:
##                          Estimate Std. Error t value Pr(>|t|)  
## (Intercept)               39.2998    24.6323   1.595   0.1181  
## tbiqol_mobility_tscore     0.6774     0.3671   1.845   0.0721 .
## tbiqol_genconcern_tscore   0.7580     0.3838   1.975   0.0549 .
## tbiqol_comm_tscore        -0.2437     0.4005  -0.608   0.5462  
## frsbe_apathy              -0.2018     0.3492  -0.578   0.5665  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 17.35 on 42 degrees of freedom
## Multiple R-squared:  0.2601, Adjusted R-squared:  0.1896 
## F-statistic: 3.691 on 4 and 42 DF,  p-value: 0.01156

ACS3 Fitness

## 
## Model selection based on AICc:
## 
##        K   AICc Delta_AICc AICcWt Cum.Wt      LL
## model8 7 459.79       0.00   0.46   0.46 -221.42
## model2 6 460.55       0.76   0.32   0.78 -223.20
## model7 8 462.71       2.93   0.11   0.89 -221.41
## model1 8 464.88       5.09   0.04   0.93 -222.49
## model5 8 464.95       5.17   0.04   0.96 -222.53
## model6 9 465.07       5.29   0.03   0.99 -221.04
## model4 6 469.04       9.26   0.00   1.00 -227.47
## model3 7 470.64      10.86   0.00   1.00 -226.88
## 
## Call:
## lm(formula = acsf_retain ~ phys_health_index + tbiqol_genconcern_tscore + 
##     tbiqol_anxiety_tscore + tbiqol_stigma_tscore + income, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -50.951 -17.189  -7.207   8.118 123.427 
## 
## Coefficients:
##                          Estimate Std. Error t value Pr(>|t|)  
## (Intercept)               46.9599    69.0538   0.680   0.5004  
## phys_health_index          0.3362     0.4663   0.721   0.4751  
## tbiqol_genconcern_tscore   0.2881     0.7446   0.387   0.7008  
## tbiqol_anxiety_tscore     -0.1669     0.7041  -0.237   0.8139  
## tbiqol_stigma_tscore      -0.7124     0.8671  -0.822   0.4162  
## income                    12.3356     6.8803   1.793   0.0806 .
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 31.96 on 40 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.1986, Adjusted R-squared:  0.09848 
## F-statistic: 1.983 on 5 and 40 DF,  p-value: 0.1021

RQ1b

Moderation

For moderation, looking at personal protective factors and environmental protective factors from original model

## 
## Call:
## lm(formula = acsg_retain ~ phys_health_index * frsbe_total, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -31.175  -8.852  -1.446  10.181  35.091 
## 
## Coefficients:
##                                 Estimate Std. Error t value Pr(>|t|)
## (Intercept)                    57.685155 102.569609   0.562    0.577
## phys_health_index               0.469664   1.030243   0.456    0.651
## frsbe_total                    -0.117970   0.914296  -0.129    0.898
## phys_health_index:frsbe_total  -0.001378   0.009340  -0.148    0.883
## 
## Residual standard error: 15.64 on 43 degrees of freedom
## Multiple R-squared:  0.2189, Adjusted R-squared:  0.1644 
## F-statistic: 4.018 on 3 and 43 DF,  p-value: 0.01319
## 
## Call:
## lm(formula = acsg_retain ~ tbiqol_genconcern_tscore * frsbe_total, 
##     data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -27.944  -9.354   1.289   7.825  32.664 
## 
## Coefficients:
##                                       Estimate Std. Error t value Pr(>|t|)
## (Intercept)                          51.480917  54.513263   0.944    0.350
## tbiqol_genconcern_tscore              1.063893   1.434613   0.742    0.462
## frsbe_total                          -0.051765   0.492276  -0.105    0.917
## tbiqol_genconcern_tscore:frsbe_total -0.002369   0.013757  -0.172    0.864
## 
## Residual standard error: 15.02 on 43 degrees of freedom
## Multiple R-squared:  0.2801, Adjusted R-squared:  0.2298 
## F-statistic: 5.575 on 3 and 43 DF,  p-value: 0.00254
## 
## Call:
## lm(formula = acsg_retain ~ emo_health_index * frsbe_total, data = df)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -31.4760  -7.4443   0.5405   9.4269  31.2846 
## 
## Coefficients:
##                                Estimate Std. Error t value Pr(>|t|)
## (Intercept)                  -32.783896 110.933886  -0.296    0.769
## emo_health_index               1.322256   1.037817   1.274    0.209
## frsbe_total                    0.803242   0.960737   0.836    0.408
## emo_health_index:frsbe_total  -0.010388   0.009178  -1.132    0.264
## 
## Residual standard error: 15.83 on 43 degrees of freedom
## Multiple R-squared:  0.1996, Adjusted R-squared:  0.1438 
## F-statistic: 3.574 on 3 and 43 DF,  p-value: 0.02148
## 
## Call:
## lm(formula = acsg_retain ~ spstotal * frsbe_total, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -32.313  -9.575   1.231   8.384  31.091 
## 
## Coefficients:
##                        Estimate Std. Error t value Pr(>|t|)
## (Intercept)           8.055e+01  1.136e+02   0.709    0.482
## spstotal              2.904e-01  1.410e+00   0.206    0.838
## frsbe_total          -2.342e-01  1.025e+00  -0.228    0.820
## spstotal:frsbe_total -3.385e-04  1.306e-02  -0.026    0.979
## 
## Residual standard error: 16 on 43 degrees of freedom
## Multiple R-squared:  0.1824, Adjusted R-squared:  0.1254 
## F-statistic: 3.198 on 3 and 43 DF,  p-value: 0.03269

There was no moderating effect of apathy on any of the predictors.

Mediation

Here we look at mediation effect of the total FrSBe scores on the personal and environmental factors used in the post hoc model (mobility, general cog functioning, anxiety, depression, and social support)

How to read: ACME = indirect effect. ADE = direct effect. ACME + ADE = total effect.

Physical Health

# Initial Model
model1 <- lm(acsg_retain ~ phys_health_index, df) # Y ~ X, DV predicted by IV - no mediation considered - total effect
summary(model1)
## 
## Call:
## lm(formula = acsg_retain ~ phys_health_index, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -31.509  -9.597  -2.509  11.759  34.771 
## 
## Coefficients:
##                   Estimate Std. Error t value Pr(>|t|)   
## (Intercept)        31.4292    16.3825   1.918  0.06141 . 
## phys_health_index   0.4640     0.1718   2.700  0.00973 **
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 16.05 on 45 degrees of freedom
## Multiple R-squared:  0.1394, Adjusted R-squared:  0.1203 
## F-statistic: 7.292 on 1 and 45 DF,  p-value: 0.009728
# Mediation paths
medmodel1 <- lm(frsbe_total ~ phys_health_index, df) # M ~ X, mediator predicted by X
outputmodel1 <- lm(acsg_retain ~ phys_health_index + frsbe_total, df) # Y ~ X + M, DV predicted by mediator, adjusting for IV

summary(medmodel1)
## 
## Call:
## lm(formula = frsbe_total ~ phys_health_index, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -33.392 -15.118  -1.668  12.312  48.040 
## 
## Coefficients:
##                   Estimate Std. Error t value Pr(>|t|)    
## (Intercept)        162.592     19.735   8.239 1.55e-10 ***
## phys_health_index   -0.572      0.207  -2.763  0.00826 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 19.33 on 45 degrees of freedom
## Multiple R-squared:  0.1451, Adjusted R-squared:  0.1261 
## F-statistic: 7.636 on 1 and 45 DF,  p-value: 0.00826
summary(outputmodel1)
## 
## Call:
## lm(formula = acsg_retain ~ phys_health_index + frsbe_total, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -31.321  -8.737  -1.367   9.834  35.352 
## 
## Coefficients:
##                   Estimate Std. Error t value Pr(>|t|)   
## (Intercept)        72.3506    25.0043   2.894   0.0059 **
## phys_health_index   0.3200     0.1791   1.787   0.0808 . 
## frsbe_total        -0.2517     0.1193  -2.110   0.0405 * 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 15.47 on 44 degrees of freedom
## Multiple R-squared:  0.2185, Adjusted R-squared:  0.183 
## F-statistic: 6.153 on 2 and 44 DF,  p-value: 0.004404
# Mediation test
mediation <- mediate(medmodel1, # Mediator model
                    outputmodel1, # Outcome model
                    boot = T, # Ask for bootstrapped confidence intervals
                    treat="phys_health_index", # Name of the x variable
                    mediator="frsbe_total" # Name of the m variable
                    )
# if you don't want bootstrap, just delete 'sims' line and set boot = F

summary(mediation)
## 
## Causal Mediation Analysis 
## 
## Nonparametric Bootstrap Confidence Intervals with the Percentile Method
## 
##                Estimate 95% CI Lower 95% CI Upper p-value   
## ACME             0.1440       0.0109         0.35   0.022 * 
## ADE              0.3200      -0.0354         0.67   0.070 . 
## Total Effect     0.4640       0.1259         0.84   0.008 **
## Prop. Mediated   0.3103       0.0246         1.12   0.030 * 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Sample Size Used: 47 
## 
## 
## Simulations: 1000
plot(mediation)

There is a significant indirect effect and an insignificant direct effect, indicating total mediation

## [1] "-0.57"
## [1] "-0.25"

#### Cognitive Health

## 
## Call:
## lm(formula = acsg_retain ~ tbiqol_genconcern_tscore, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -32.004 -10.422   0.877   9.549  32.010 
## 
## Coefficients:
##                          Estimate Std. Error t value Pr(>|t|)    
## (Intercept)               38.9504     9.3250   4.177 0.000134 ***
## tbiqol_genconcern_tscore   1.0047     0.2513   3.998 0.000235 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 14.86 on 45 degrees of freedom
## Multiple R-squared:  0.2621, Adjusted R-squared:  0.2457 
## F-statistic: 15.99 on 1 and 45 DF,  p-value: 0.0002345
## 
## Call:
## lm(formula = frsbe_total ~ tbiqol_genconcern_tscore, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -34.149  -9.953  -0.151  11.214  40.165 
## 
## Coefficients:
##                          Estimate Std. Error t value Pr(>|t|)    
## (Intercept)              157.8158    10.7369  14.699  < 2e-16 ***
## tbiqol_genconcern_tscore  -1.3631     0.2893  -4.711 2.39e-05 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 17.11 on 45 degrees of freedom
## Multiple R-squared:  0.3303, Adjusted R-squared:  0.3154 
## F-statistic:  22.2 on 1 and 45 DF,  p-value: 2.39e-05
## 
## Call:
## lm(formula = acsg_retain ~ tbiqol_genconcern_tscore + frsbe_total, 
##     data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -28.123  -9.765   1.326   7.881  32.669 
## 
## Coefficients:
##                          Estimate Std. Error t value Pr(>|t|)  
## (Intercept)               60.0145    22.4434   2.674   0.0105 *
## tbiqol_genconcern_tscore   0.8227     0.3068   2.681   0.0103 *
## frsbe_total               -0.1335     0.1294  -1.032   0.3079  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 14.85 on 44 degrees of freedom
## Multiple R-squared:  0.2796, Adjusted R-squared:  0.2468 
## F-statistic: 8.537 on 2 and 44 DF,  p-value: 0.0007366
## 
## Causal Mediation Analysis 
## 
## Nonparametric Bootstrap Confidence Intervals with the Percentile Method
## 
##                Estimate 95% CI Lower 95% CI Upper p-value    
## ACME              0.182       -0.118         0.54   0.238    
## ADE               0.823        0.265         1.31   0.004 ** 
## Total Effect      1.005        0.532         1.49  <2e-16 ***
## Prop. Mediated    0.181       -0.131         0.61   0.238    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Sample Size Used: 47 
## 
## 
## Simulations: 1000

There is no significant indirect effect. No mediation

Emotional Health

## 
## Call:
## lm(formula = acsg_retain ~ emo_health_index, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -34.933  -9.024  -0.002   9.921  32.264 
## 
## Coefficients:
##                  Estimate Std. Error t value Pr(>|t|)  
## (Intercept)       37.7676    17.3514   2.177   0.0348 *
## emo_health_index   0.3777     0.1733   2.179   0.0346 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 16.45 on 45 degrees of freedom
## Multiple R-squared:  0.09544,    Adjusted R-squared:  0.07534 
## F-statistic: 4.748 on 1 and 45 DF,  p-value: 0.03461
## 
## Call:
## lm(formula = frsbe_total ~ emo_health_index, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -28.203 -12.814  -2.400   7.978  48.254 
## 
## Coefficients:
##                  Estimate Std. Error t value Pr(>|t|)    
## (Intercept)      184.3468    18.8749   9.767 1.08e-12 ***
## emo_health_index  -0.7638     0.1885  -4.051 0.000199 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 17.9 on 45 degrees of freedom
## Multiple R-squared:  0.2672, Adjusted R-squared:  0.251 
## F-statistic: 16.41 on 1 and 45 DF,  p-value: 0.0001989
## 
## Call:
## lm(formula = acsg_retain ~ emo_health_index + frsbe_total, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -32.767  -8.418  -0.429   8.742  32.342 
## 
## Coefficients:
##                  Estimate Std. Error t value Pr(>|t|)   
## (Intercept)       88.2618    29.5862   2.983  0.00464 **
## emo_health_index   0.1685     0.1955   0.862  0.39346   
## frsbe_total       -0.2739     0.1323  -2.070  0.04431 * 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 15.88 on 44 degrees of freedom
## Multiple R-squared:  0.1757, Adjusted R-squared:  0.1383 
## F-statistic: 4.691 on 2 and 44 DF,  p-value: 0.01423
## 
## Causal Mediation Analysis 
## 
## Nonparametric Bootstrap Confidence Intervals with the Percentile Method
## 
##                Estimate 95% CI Lower 95% CI Upper p-value  
## ACME             0.2092       0.0171         0.43   0.040 *
## ADE              0.1685      -0.2411         0.58   0.370  
## Total Effect     0.3777       0.0470         0.72   0.036 *
## Prop. Mediated   0.5540      -0.1432         2.81   0.076 .
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Sample Size Used: 47 
## 
## 
## Simulations: 1000

There is no significant indirect effect indicating no mediation

## [1] "-0.76"
## [1] "-0.27"

#### SPS

## 
## Call:
## lm(formula = acsg_retain ~ spstotal, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -33.673  -9.945  -0.095   8.838  32.349 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)  
## (Intercept)  35.4478    17.0229   2.082   0.0430 *
## spstotal      0.5027     0.2131   2.359   0.0227 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 16.32 on 45 degrees of freedom
## Multiple R-squared:  0.1101, Adjusted R-squared:  0.09029 
## F-statistic: 5.566 on 1 and 45 DF,  p-value: 0.02272
## 
## Call:
## lm(formula = frsbe_total ~ spstotal, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -39.247 -14.035  -0.331  10.358  35.999 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept) 183.9965    18.6242   9.879 7.59e-13 ***
## spstotal     -0.9529     0.2331  -4.088 0.000177 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 17.86 on 45 degrees of freedom
## Multiple R-squared:  0.2708, Adjusted R-squared:  0.2546 
## F-statistic: 16.71 on 1 and 45 DF,  p-value: 0.0001774
## 
## Call:
## lm(formula = acsg_retain ~ spstotal + frsbe_total, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -32.358  -9.535   1.281   8.379  31.082 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)   
## (Intercept)  83.3909    29.3743   2.839  0.00683 **
## spstotal      0.2544     0.2419   1.052  0.29865   
## frsbe_total  -0.2606     0.1321  -1.973  0.05482 . 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 15.82 on 44 degrees of freedom
## Multiple R-squared:  0.1824, Adjusted R-squared:  0.1452 
## F-statistic: 4.908 on 2 and 44 DF,  p-value: 0.01191
## 
## Causal Mediation Analysis 
## 
## Nonparametric Bootstrap Confidence Intervals with the Percentile Method
## 
##                Estimate 95% CI Lower 95% CI Upper p-value  
## ACME             0.2483       0.0383         0.56   0.018 *
## ADE              0.2544      -0.2218         0.70   0.304  
## Total Effect     0.5027       0.0624         0.93   0.030 *
## Prop. Mediated   0.4939       0.0212         2.33   0.048 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Sample Size Used: 47 
## 
## 
## Simulations: 1000

There is a significant indirect effect and insignificant direct effect indicating full mediation

## [1] "-0.95"
## [1] "-0.26"

Post Hoc Mediation

In this set of mediation analysis, we look specifically at apathy as a mediator for the ACS social re-engagement outcome

## 
## Call:
## lm(formula = acss_retain ~ tbiqol_depression_tscore, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -40.638 -10.314   0.074  11.575  34.999 
## 
## Coefficients:
##                          Estimate Std. Error t value Pr(>|t|)    
## (Intercept)              116.4764    15.1471   7.690 9.79e-10 ***
## tbiqol_depression_tscore  -0.8087     0.2761  -2.929  0.00532 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 18.13 on 45 degrees of freedom
## Multiple R-squared:  0.1602, Adjusted R-squared:  0.1415 
## F-statistic: 8.581 on 1 and 45 DF,  p-value: 0.005317
## 
## Call:
## lm(formula = frsbe_apathy ~ tbiqol_depression_tscore, data = df)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -12.1753  -6.2877  -0.3744   5.1128  20.9676 
## 
## Coefficients:
##                          Estimate Std. Error t value Pr(>|t|)    
## (Intercept)               10.5876     6.3297   1.673 0.101327    
## tbiqol_depression_tscore   0.4215     0.1154   3.654 0.000671 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 7.578 on 45 degrees of freedom
## Multiple R-squared:  0.2288, Adjusted R-squared:  0.2117 
## F-statistic: 13.35 on 1 and 45 DF,  p-value: 0.0006714
## 
## Call:
## lm(formula = acss_retain ~ tbiqol_depression_tscore + frsbe_apathy, 
##     data = df)
## 
## Residuals:
##    Min     1Q Median     3Q    Max 
## -36.30 -10.39  -1.80  16.06  26.89 
## 
## Coefficients:
##                          Estimate Std. Error t value Pr(>|t|)    
## (Intercept)              124.6480    14.9436   8.341 1.31e-10 ***
## tbiqol_depression_tscore  -0.4833     0.3009  -1.606   0.1154    
## frsbe_apathy              -0.7718     0.3415  -2.260   0.0288 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 17.36 on 44 degrees of freedom
## Multiple R-squared:  0.2475, Adjusted R-squared:  0.2133 
## F-statistic: 7.236 on 2 and 44 DF,  p-value: 0.001919
## 
## Causal Mediation Analysis 
## 
## Nonparametric Bootstrap Confidence Intervals with the Percentile Method
## 
##                Estimate 95% CI Lower 95% CI Upper p-value    
## ACME            -0.3254      -0.7283        -0.06   0.016 *  
## ADE             -0.4833      -1.1235         0.14   0.136    
## Total Effect    -0.8087      -1.3168        -0.32  <2e-16 ***
## Prop. Mediated   0.4023       0.0614         1.42   0.016 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Sample Size Used: 47 
## 
## 
## Simulations: 1000

## [1] "0.42"
## [1] "-0.77"

## 
## Call:
## lm(formula = acss_retain ~ tbiqol_fatigue_tscore, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -36.178 -10.715   1.675  14.500  40.993 
## 
## Coefficients:
##                       Estimate Std. Error t value Pr(>|t|)    
## (Intercept)           129.0292    17.0969   7.547 1.59e-09 ***
## tbiqol_fatigue_tscore  -1.0210     0.3068  -3.328  0.00175 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 17.73 on 45 degrees of freedom
## Multiple R-squared:  0.1975, Adjusted R-squared:  0.1797 
## F-statistic: 11.07 on 1 and 45 DF,  p-value: 0.001751
## 
## Call:
## lm(formula = frsbe_apathy ~ tbiqol_fatigue_tscore, data = df)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -12.8790  -4.8855  -0.3758   4.5242  15.9984 
## 
## Coefficients:
##                       Estimate Std. Error t value Pr(>|t|)    
## (Intercept)             4.9308     7.1335   0.691 0.492975    
## tbiqol_fatigue_tscore   0.5161     0.1280   4.032 0.000211 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 7.396 on 45 degrees of freedom
## Multiple R-squared:  0.2654, Adjusted R-squared:  0.2491 
## F-statistic: 16.26 on 1 and 45 DF,  p-value: 0.0002112
## 
## Call:
## lm(formula = acss_retain ~ tbiqol_fatigue_tscore + frsbe_apathy, 
##     data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -35.271 -13.323  -0.586  13.963  40.058 
## 
## Coefficients:
##                       Estimate Std. Error t value Pr(>|t|)    
## (Intercept)           132.4470    16.6389   7.960  4.6e-10 ***
## tbiqol_fatigue_tscore  -0.6633     0.3465  -1.914   0.0621 .  
## frsbe_apathy           -0.6931     0.3459  -2.004   0.0512 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 17.16 on 44 degrees of freedom
## Multiple R-squared:  0.2646, Adjusted R-squared:  0.2312 
## F-statistic: 7.917 on 2 and 44 DF,  p-value: 0.001157
## 
## Causal Mediation Analysis 
## 
## Nonparametric Bootstrap Confidence Intervals with the Percentile Method
## 
##                Estimate 95% CI Lower 95% CI Upper p-value    
## ACME            -0.3578      -0.8107        -0.11   0.004 ** 
## ADE             -0.6633      -1.3189         0.04   0.072 .  
## Total Effect    -1.0210      -1.6485        -0.44  <2e-16 ***
## Prop. Mediated   0.3504       0.0928         1.10   0.004 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Sample Size Used: 47 
## 
## 
## Simulations: 1000

## [1] "0.52"
## [1] "-0.69"

## 
## Call:
## lm(formula = acss_retain ~ tbiqol_genconcern_tscore, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -42.828 -13.186   4.512  11.990  26.980 
## 
## Coefficients:
##                          Estimate Std. Error t value Pr(>|t|)    
## (Intercept)               31.4994    10.6832   2.949 0.005049 ** 
## tbiqol_genconcern_tscore   1.1439     0.2879   3.974 0.000253 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 17.02 on 45 degrees of freedom
## Multiple R-squared:  0.2597, Adjusted R-squared:  0.2433 
## F-statistic: 15.79 on 1 and 45 DF,  p-value: 0.0002532
## 
## Call:
## lm(formula = frsbe_apathy ~ tbiqol_genconcern_tscore, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -19.282  -5.537   0.206   4.894  14.257 
## 
## Coefficients:
##                          Estimate Std. Error t value Pr(>|t|)    
## (Intercept)               48.7536     4.8739  10.003 5.15e-13 ***
## tbiqol_genconcern_tscore  -0.4264     0.1313  -3.247  0.00221 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 7.767 on 45 degrees of freedom
## Multiple R-squared:  0.1898, Adjusted R-squared:  0.1718 
## F-statistic: 10.54 on 1 and 45 DF,  p-value: 0.002207
## 
## Call:
## lm(formula = acss_retain ~ tbiqol_genconcern_tscore + frsbe_apathy, 
##     data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -35.659 -11.474  -0.009  13.358  28.851 
## 
## Coefficients:
##                          Estimate Std. Error t value Pr(>|t|)   
## (Intercept)               63.0926    18.5305   3.405  0.00142 **
## tbiqol_genconcern_tscore   0.8676     0.3090   2.808  0.00741 **
## frsbe_apathy              -0.6480     0.3157  -2.053  0.04607 * 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 16.45 on 44 degrees of freedom
## Multiple R-squared:  0.3244, Adjusted R-squared:  0.2937 
## F-statistic: 10.57 on 2 and 44 DF,  p-value: 0.0001789
## 
## Causal Mediation Analysis 
## 
## Nonparametric Bootstrap Confidence Intervals with the Percentile Method
## 
##                Estimate 95% CI Lower 95% CI Upper p-value    
## ACME            0.27634      0.00381         0.59   0.048 *  
## ADE             0.86757      0.32064         1.40  <2e-16 ***
## Total Effect    1.14391      0.63978         1.63  <2e-16 ***
## Prop. Mediated  0.24158      0.00720         0.59   0.048 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Sample Size Used: 47 
## 
## 
## Simulations: 1000

## [1] "-0.43"
## [1] "-0.65"

RQ2: Time Since Injury

  1. How does time since injury influence resiliency?
  1. Is time after injury associated with an individual’s ability to re-engage in meaningful activities?
  2. To what extent do the relationships between protective factors and self-regulatory processes with resiliency-related outcomes change with time since sustaining TBI? Hypothesis: Time since injury will be associated with resiliency, leading to different resiliency-related outcomes

To answer these questions, first look at descriptive statistics, then regression model with time since injury included, lastly, investigate what, if any, role time since injury has on protective factors and self-regulation

Descriptives

##             Min. 1st Qu. Median Mean 3rd Qu. Max.
## time_injury    1    3.25      7 8.96      11   30
## [1] 8.957447
## [1] 7.079387

Correlation ACS

Regression Analysis

Global ACS

First, looking just at the relationship between time since injury and re-engagement while controlling for age.

## 
## Call:
## lm(formula = acsg_retain ~ age_current + time_injury, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -38.997  -9.010  -0.178   9.243  40.142 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  83.3414     8.1920  10.174 3.92e-13 ***
## age_current  -0.2991     0.1675  -1.786   0.0810 .  
## time_injury   0.6458     0.3495   1.848   0.0714 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 16.48 on 44 degrees of freedom
## Multiple R-squared:  0.1122, Adjusted R-squared:  0.07188 
## F-statistic: 2.781 on 2 and 44 DF,  p-value: 0.07288

Controlling for age, there is no significant relationship between time since injury and re-engagement

Social ACS3

## 
## Call:
## lm(formula = acss_retain ~ age_current + time_injury, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -43.803 -11.929   4.035  14.187  28.866 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  79.9908     9.2439   8.653 4.75e-11 ***
## age_current  -0.3273     0.1890  -1.732   0.0902 .  
## time_injury   0.8955     0.3944   2.271   0.0281 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 18.6 on 44 degrees of freedom
## Multiple R-squared:  0.136,  Adjusted R-squared:  0.09669 
## F-statistic: 3.462 on 2 and 44 DF,  p-value: 0.04015

IADL ACS3

## 
## Call:
## lm(formula = acsi_retain ~ age_current + time_injury, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -33.423 -15.564   0.161  16.149  30.936 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  86.0488     8.9658   9.597 2.34e-12 ***
## age_current  -0.2444     0.1833  -1.333    0.189    
## time_injury   0.6126     0.3825   1.602    0.116    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 18.04 on 44 degrees of freedom
## Multiple R-squared:  0.077,  Adjusted R-squared:  0.03504 
## F-statistic: 1.835 on 2 and 44 DF,  p-value: 0.1716

Leisure ACS3

## 
## Call:
## lm(formula = acsl_retain ~ age_current + time_injury, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -46.198 -10.286  -1.274  13.458  52.180 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  89.1879     9.3269   9.562 2.61e-12 ***
## age_current  -0.3264     0.1907  -1.712   0.0939 .  
## time_injury   0.6217     0.3979   1.562   0.1254    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 18.77 on 44 degrees of freedom
## Multiple R-squared:  0.09329,    Adjusted R-squared:  0.05208 
## F-statistic: 2.264 on 2 and 44 DF,  p-value: 0.1159

Fitness ACS3

## 
## Call:
## lm(formula = acsf_retain ~ age_current + time_injury, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -50.303 -16.695  -4.872  10.540 127.414 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  73.9164    16.3563   4.519 4.63e-05 ***
## age_current  -0.3726     0.3344  -1.114    0.271    
## time_injury   1.0721     0.6979   1.536    0.132    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 32.91 on 44 degrees of freedom
## Multiple R-squared:  0.06517,    Adjusted R-squared:  0.02267 
## F-statistic: 1.534 on 2 and 44 DF,  p-value: 0.2271

Hierarchical Regression: Step 4

## 
## Call:
## lm(formula = acsg_retain ~ age_current + phys_health_index + 
##     tbiqol_genconcern_tscore + emo_health_index + spstotal + 
##     frsbe_total + time_injury, data = df)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -27.214  -7.610  -2.886   7.239  33.115 
## 
## Coefficients:
##                          Estimate Std. Error t value Pr(>|t|)  
## (Intercept)               46.0409    37.8837   1.215   0.2315  
## age_current               -0.1030     0.1597  -0.645   0.5227  
## phys_health_index          0.1833     0.2235   0.820   0.4171  
## tbiqol_genconcern_tscore   0.6243     0.3865   1.615   0.1143  
## emo_health_index          -0.1721     0.2224  -0.774   0.4437  
## spstotal                   0.2642     0.2486   1.063   0.2944  
## frsbe_total               -0.1579     0.1374  -1.149   0.2575  
## time_injury                0.8321     0.3105   2.680   0.0107 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 14.36 on 39 degrees of freedom
## Multiple R-squared:  0.4029, Adjusted R-squared:  0.2957 
## F-statistic:  3.76 on 7 and 39 DF,  p-value: 0.003321
#Nested Model Comparison
anova(step3, step4)
## Analysis of Variance Table
## 
## Model 1: acsg_retain ~ age_current + phys_health_index + tbiqol_genconcern_tscore + 
##     emo_health_index + spstotal + frsbe_total
## Model 2: acsg_retain ~ age_current + phys_health_index + tbiqol_genconcern_tscore + 
##     emo_health_index + spstotal + frsbe_total + time_injury
##   Res.Df    RSS Df Sum of Sq      F  Pr(>F)  
## 1     40 9522.7                              
## 2     39 8041.5  1    1481.2 7.1838 0.01072 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#change in R-squared
summary(step4)$r.squared - summary(step3)$r.squared
## [1] 0.1099826

POST HOC: Groups

Because there was no significant relationship between time since injury and outcomes BUT it was a significant predictor in the model, I seperated time since recovery into three groups to examine relationships further. Early = 6mo-3years Mid = 3 to 10 years Later = >10 years

1= everything <=3 and 3 is everything >=10

Characteristic Early, N = 121 Mid, N = 141 Later, N = 211
Age (years) 39 (14) 50 (17) 48 (13)
Employment status


    0 5 (42%) 11 (79%) 12 (57%)
    1 7 (58%) 3 (21%) 9 (43%)
Substance use score 2.00 (1.71) 3.21 (3.53) 2.81 (3.16)
Severity of Injury


    2 5 (42%) 7 (50%) 7 (33%)
    3 7 (58%) 7 (50%) 14 (67%)
Global ACS 68 (16) 73 (18) 80 (16)
Social ACS 67 (20) 69 (14) 79 (22)
IADL ACS 72 (19) 79 (20) 85 (16)
Leisure ACS 71 (19) 79 (21) 85 (17)
Fitness ACS 58 (22) 61 (31) 74 (40)
1 Mean (SD); n (%)
## 
## Early   Mid Later 
##    12    14    21

We see the counts of # participants in each group

Global

Global ACS3 scores (ie, global re-engagement scores)

## # A tibble: 3 × 4
##   time_injury_ex count  mean    sd
##   <fct>          <int> <dbl> <dbl>
## 1 Early             12  68.3  16.0
## 2 Mid               14  73.4  18.3
## 3 Later             21  80.4  16.0

##                Df Sum Sq Mean Sq F value Pr(>F)
## time_injury_ex  2   1177   588.5   2.107  0.134
## Residuals      44  12291   279.3
## 
## Call:
## lm(formula = acsg_retain ~ age_current + time_injury_ex, data = df_time3)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -34.555 -10.753   0.697   9.204  38.098 
## 
## Coefficients:
##                     Estimate Std. Error t value Pr(>|t|)    
## (Intercept)          81.7469     8.0927  10.101 6.35e-13 ***
## age_current          -0.3425     0.1689  -2.028   0.0488 *  
## time_injury_exMid     8.7584     6.6156   1.324   0.1925    
## time_injury_exLater  15.1870     6.0465   2.512   0.0159 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 16.15 on 43 degrees of freedom
## Multiple R-squared:  0.1671, Adjusted R-squared:  0.1089 
## F-statistic: 2.875 on 3 and 43 DF,  p-value: 0.04708

Controlling for age, we see a significant relationship between time since injury and global re engagement- specifically between early and later recovery

Social

social ACS3 scores (ie, social re-engagement scores)

## # A tibble: 3 × 4
##   time_injury_ex count  mean    sd
##   <fct>          <int> <dbl> <dbl>
## 1 Early             12  66.7  19.8
## 2 Mid               14  69.4  14.2
## 3 Later             21  78.5  21.7

##                Df Sum Sq Mean Sq F value Pr(>F)
## time_injury_ex  2   1299   649.3    1.75  0.186
## Residuals      44  16321   370.9
## 
## Call:
## lm(formula = acss_retain ~ age_current + time_injury_ex, data = df_time3)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -40.765 -11.897   4.568  16.424  27.008 
## 
## Coefficients:
##                     Estimate Std. Error t value Pr(>|t|)    
## (Intercept)          79.8136     9.4482   8.448 1.11e-10 ***
## age_current          -0.3357     0.1972  -1.702   0.0959 .  
## time_injury_exMid     6.4223     7.7237   0.832   0.4103    
## time_injury_exLater  14.9341     7.0592   2.116   0.0402 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 18.86 on 43 degrees of freedom
## Multiple R-squared:  0.1322, Adjusted R-squared:  0.07165 
## F-statistic: 2.183 on 3 and 43 DF,  p-value: 0.1038

IADL

IADL ACS3 scores (ie, IADL re-engagement scores)

## # A tibble: 3 × 4
##   time_injury_ex count  mean    sd
##   <fct>          <int> <dbl> <dbl>
## 1 Early             12  72.2  18.8
## 2 Mid               14  79.1  20.2
## 3 Later             21  85.4  15.8

##                Df Sum Sq Mean Sq F value Pr(>F)
## time_injury_ex  2   1364   682.1   2.121  0.132
## Residuals      44  14153   321.6
## 
## Call:
## lm(formula = acsi_retain ~ age_current + time_injury_ex, data = df_time3)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -26.6008 -14.5223  -0.4196  14.5986  30.4952 
## 
## Coefficients:
##                     Estimate Std. Error t value Pr(>|t|)    
## (Intercept)          83.9693     8.8190   9.521 3.74e-12 ***
## age_current          -0.3013     0.1840  -1.637   0.1088    
## time_injury_exMid    10.2623     7.2093   1.423   0.1618    
## time_injury_exLater  16.0242     6.5891   2.432   0.0193 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 17.6 on 43 degrees of freedom
## Multiple R-squared:  0.1414, Adjusted R-squared:  0.08155 
## F-statistic: 2.361 on 3 and 43 DF,  p-value: 0.08459

We see a significant difference between early and late groups with IADl engagement when controlling for age

Leisure

Leisure ACS3 scores (ie, leisure re-engagement scores)

## # A tibble: 3 × 4
##   time_injury_ex count  mean    sd
##   <fct>          <int> <dbl> <dbl>
## 1 Early             12  71.2  19.3
## 2 Mid               14  78.8  21.0
## 3 Later             21  84.9  17.0

##                Df Sum Sq Mean Sq F value Pr(>F)
## time_injury_ex  2   1426   713.2   2.003  0.147
## Residuals      44  15667   356.1
## 
## Call:
## lm(formula = acsl_retain ~ age_current + time_injury_ex, data = df_time3)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -40.426 -13.860   2.239  10.586  49.506 
## 
## Coefficients:
##                     Estimate Std. Error t value Pr(>|t|)    
## (Intercept)          86.7213     9.1178   9.511 3.86e-12 ***
## age_current          -0.3950     0.1903  -2.076   0.0439 *  
## time_injury_exMid    11.8432     7.4536   1.589   0.1194    
## time_injury_exLater  17.2281     6.8124   2.529   0.0152 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 18.2 on 43 degrees of freedom
## Multiple R-squared:  0.1669, Adjusted R-squared:  0.1088 
## F-statistic: 2.872 on 3 and 43 DF,  p-value: 0.0472

We see a significant difference between early and late groups with Leisure engagement when controlling for age

Fitness

Fitness ACS3 scores (ie, fitness re-engagement scores)

## # A tibble: 3 × 4
##   time_injury_ex count  mean    sd
##   <fct>          <int> <dbl> <dbl>
## 1 Early             12  58.1  21.7
## 2 Mid               14  61.4  30.6
## 3 Later             21  74.0  39.5

##                Df Sum Sq Mean Sq F value Pr(>F)
## time_injury_ex  2   2412    1206   1.092  0.344
## Residuals      44  48575    1104
## 
## Call:
## lm(formula = acsf_retain ~ age_current + time_injury_ex, data = df_time3)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -57.420 -18.947  -5.366  10.466 132.029 
## 
## Coefficients:
##                     Estimate Std. Error t value Pr(>|t|)    
## (Intercept)          73.2745    16.5997   4.414 6.71e-05 ***
## age_current          -0.3879     0.3464  -1.120    0.269    
## time_injury_exMid     7.5033    13.5700   0.553    0.583    
## time_injury_exLater  19.5197    12.4025   1.574    0.123    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 33.13 on 43 degrees of freedom
## Multiple R-squared:  0.0743, Adjusted R-squared:  0.009714 
## F-statistic:  1.15 on 3 and 43 DF,  p-value: 0.3397

Impact on Predictive Variables

This is exploratory and post hoc analysis- likely unable to report to avoid p-hacking, more for information gathering

Correlations

Self Regulation

##                Df Sum Sq Mean Sq F value Pr(>F)
## time_injury_ex  2     35    17.6   0.039  0.961
## Residuals      44  19638   446.3

Physical Health

##                Df Sum Sq Mean Sq F value Pr(>F)
## time_injury_ex  2    190   94.92   0.489  0.616
## Residuals      44   8533  193.93

Emotional Health

##                Df Sum Sq Mean Sq F value Pr(>F)
## time_injury_ex  2    322   160.8   0.814   0.45
## Residuals      44   8690   197.5

General Cognition

##                Df Sum Sq Mean Sq F value Pr(>F)
## time_injury_ex  2     24   12.14   0.154  0.858
## Residuals      44   3473   78.94

Social Support

##                Df Sum Sq Mean Sq F value Pr(>F)
## time_injury_ex  2     24   12.14   0.154  0.858
## Residuals      44   3473   78.94

Cont Exploratory Analysis

Below is the comparison of t-values for step 3 and step 4 to see if the inclusion of time since injury significantly changed the contribution of each variable in the model. (not sure this is an appropriate way to report in paper, but wanted to see…)

##                   Variable     z_score    p_value
## 1              age_current  1.87537056 0.06074176
## 2        phys_health_index -0.65610266 0.51175808
## 3         emo_health_index  0.65707013 0.51113582
## 4 tbiqol_genconcern_tscore  0.03873711 0.96909999
## 5                 spstotal -1.06289614 0.28782905
## 6              frsbe_total  1.51617622 0.12947480

Selection for gift cards

Create a sequence from 1 to 44 numbers <- 1:46

Randomly select 8 numbers selected_numbers <- sample(numbers, 8)

Format the selected numbers with leading zeros formatted_numbers <- sprintf(“%03d”, selected_numbers)

Print the selected numbers print(formatted_numbers)

“020” “007” “006” “019” “024” “041” “013” “037”